AFR%Error doesn't agree with STFT's

[FSAE_Junkie]

I'll do my best to explain a problem that has been nagging at me. I have a 2006 M6 GTO. I just recently install some Kooks LT's, Vararam CAI, Ported Intake, and LC-1 Wideband. I have been following the guide in the HPT help file and have yet to get past the VE tuning stage.

In the help file, it says to turn off LTFT's and STFT's. I did this, and drove around logging AFR%Error and then applied this to my VE table. I did several iterations of this to the point where AFR%Error hoverered around -1 or -2 % error. Did the same thing with WOT areas of the VE table. I have also set PE to be a 12.6 AFR for now.


Later, I turned on STFT's just to see what they were doing, and this is where my question comes in. On several logs, my AFR%Error at cruise/idle shows -1 or -2. However, STFT's in the same cells show +4 or +5. This is about 6% error between the two. Does this mean my WB is off by ~6%? This would mean all my WOT AFR's are actually 13.4's and not 12.6's. For this reason, I've had to stop tuning until I get this figured out.

Also, I checked for voltage offsets (even though they are grounded to the same place) and Logworks LC-1 readings matched dead nuts the HPT readings, so I don't think voltage offset is a problem.

Is it common for there to be such a discrepancy? What should I trust more? If I trust the LC-1, then when I go back to Closed Loop, the STFT's are just going to be sitting at ~6%. If I trust, the STFT's, then whats the point of a wideband and how can I trust it at WOT?

[FSAE_Junkie]

Figured I'd post my tune and Log.



Another thought on this. The STFT's are the fuel currently being added, not what the PCM 'thinks' it should add in the future right? So does this mean that the PCM is 'adding' 4 to 5% more fuel via STFT's and the LC-1 is seeing this and saying 1-2% needs to be cut out?

This makes much more sense now if that is correct. That would mean that the Wideband vs STFT error is only 2-3% which seems much more reasonable than 6%.

Thoughts?

[slammin86]

Does your wideband have a stead state calibration mode that you can use to confirm that the voltage offset is good?

On my aem, when I unhook the sensor from it, it reads a steady 14.8. I then adjusted the offset in hp tuners till it matched. I ended up using 9.7 inside the software.

[JimMueller]

TTT - I'm experiencing the same discrepancy and would also know which to trust.

Edit - thinking about it some more after re-reading the thread... STFT's are how much fuel the PCM is adding based upon feedback from the O2's. So if your STFT's are 4-5%, then that 4-5% has already been added and your LC-1 would be indicating how much less/more fuel it thinks is necessary on top of the change made by your O2's.

In my case, there are some cells which show a double-digit discrepancy between STFT and WB.

[FSAE_Junkie]

Slammin, because I have HPT ground connected to the same terminal as the LC-1 ground the measured offset between them is 0.

TTT - I'm experiencing the same discrepancy and would also know which to trust.

Edit - thinking about it some more after re-reading the thread... STFT's are how much fuel the PCM is adding based upon feedback from the O2's. So if your STFT's are 4-5%, then that 4-5% has already been added and your LC-1 would be indicating how much less/more fuel it thinks is necessary on top of the change made by your O2's.

In my case, there are some cells which show a double-digit discrepancy between STFT and WB.

This is exactly what I have come to the conlusion of. Thus, you use the STFT's to tune the idle/cruise areas because this is ultimately what is going to affect the fueling. You want them to be 0. I am fortunate enought because my WB error only shows about 1-2% difference which I have heard is acceptable.

[5_Liter_Eater]

Create a histo that plots the AFR into one cell to get the average AFR. Is the average closed loop AFR near 14.7:1 according to the WB? If the average closed loop AFR is lean, more like 15.5:1, it often times indicates an air leak near or at the WB bung. Narrow bands are very accurate at 14.7:1 so they should be trusted over the WB.

[madvette08]

We have seen a error of up to 5% during part throttle tuning, when comparing the wide band to the narrow bands. That is why we recommend using the narrow bands to tune part throttle. Use the VE - LTFTs histogram for the VE table and you will have to create a histogram to tune the MAF -LTFTs. I have attached a screen shot on how to make the MAF- LTFT hisotgram.

Ian

[JimMueller]

What's this MAF thing you mention? Is that the term native southerners use to talk about arithmetic? (I'm MAFless)

You want to use the average STFT for both banks, right? And if you use STFT to tune VE through 4000RPM... you're using STFT's to tune VE all the way to 100+ kPa? Don't you want to use the WB at some kPa level and above? How do you integrate the results from the STFT's to the WB areas?

[ct06gto]

It looks like the standard ve vs stft does average both banks.


Trying my best to stay in a steady state the highest kpa cell i hit at 4000 rpm was 49 kpa.

[Z06SUPERCHARGED]

Create a histo that plots the AFR into one cell to get the average AFR. Is the average closed loop AFR near 14.7:1 according to the WB? If the average closed loop AFR is lean, more like 15.5:1, it often times indicates an air leak near or at the WB bung. Narrow bands are very accurate at 14.7:1 so they should be trusted over the WB.

Do you think NB's are accurate all the way to 4000 more so over the WB?

Im going back and forth every week tuning both ways, once using WB and oince using NB's and some time I think one is better than the other but really cant decide where to leave it.

[5_Liter_Eater]

As long as you're in closed loop they are accurate. If your average WB reading in closed loop is not ~14.7 then it points to a problem with the WB. If it's reading leaner then it's likely an air leak around the WB bung as mentioned above.

[NassyVette]

This may be a stupid question, but if the narrow band O2's and MAF work so well at keeping the car at stoich, why bother narrowing the VE to within -2? I get that it helps with transitions where PE takes over etc, but if you are going back to closed loop and letting the computer take over it seems like a waste of time to get it zeroed in, and would seem to only be worthwhile on a true SD tune?

[5_Liter_Eater]

Because if your fuel trims are positive then that gets added to the PE. IE: if the LTFT cell you left closed loop in was a +5 then 5% gets added to your PE. So 1.131 (13:1) becomes 1.188 (12.37:1).

This is why I and others disable LTFT's and only use STFTs. They don't get added to PE if they are positive.

[NassyVette]

Because if your fuel trims are positive then that gets added to the PE. IE: if the LTFT cell you left closed loop in was a +5 then 5% gets added to your PE. So 1.131 (13:1) becomes 1.188 (12.37:1).

This is why I and others disable LTFT's and only use STFTs. They don't get added to PE if they are positive.

That makes sense... are you permanently disabling LTFT's after VE tuning? In other words, tune the VE, tune MAF, re-enable closed loop and disable LTFT's? Then finish fine tuning with the very accurate NB o2's using stft's? Is that correct?

Slightly off topic, but what would you recommend setting the PE multiplier at when tuning the VE in open loop? Right now I have a commanded AFR of 12.6. My new motor is 11.5:1 CR and I don't want to creep too high on the AFR till I'm on a dyno. Does the commanded afr of 12.6 make sense or should I command 12.8-13.0 and monitor KR and AFR for lean spikes etc and adjust accordingly?

[5_Liter_Eater]

Yes, I disable them entirely. So I'll OL WB tune, then re-enable closed loop (STFTs only) and double check the STFT trims. Then I leave the LTFTs disabled for every day driving.

Leave PE what you want your WOT AFR to be. Artificially increase the entire VE table by 5-10% so you are sure to start off on the rich side.